Shareholder rights law firm Johnson Fistel, LLP ( is investigating whether Amplitude, Inc. (NASDAQ:AMPL), any of its executive officers or others

AMPL SHAREHOLDER ALERT: Johnson Fistel Investigates Amplitude

Despite recent volatility, U.S. equities markets have had a much better year than expected in 2023. The S&P 500 and Nasdaq are only returning 9.8% and 25.5% since the start of the year. Geopolitical events that could move the world in different directions and soaring treasury yields were to blame for the market sell-off last week.

3 Struggling Stocks to Sell in November Before They Implode

InvestorPlace - Stock Market News, Stock Advice & Trading Tips Despite recent volatility, U.S. equities markets have had a much better year

These WWE tag teams have crafted an unforgettable legacy in the promotion, supported by some truly amazing finishers.

10 Best WWE Tag Teams Ever, Ranked By Their Finisher

Figure 3. Silicon Creations DP RX JTOL vs DP mask DP Main Link Source Device Electrical Requirements: The DP Source device must consist of a transmitter that can transmit the required data rate at specified amplitude levels. The TX must also support ...

High Speed, Low Power and Flexibility Drive DisplayPort's Increasing Popularity

To determine the effect of melanin content on conventional amplitude-based PA images, as well as the impact of SLSC beamforming, they calculated the signal-to-noise ratio (SNR) of the radial artery running through the forearm and the level of surrounding ...

High-quality diagnostic imaging – for all skin tones

The Endless universe sprang to life from the strategic wranglings of Endless Space, a 4X turn-based strategy game that frankly couldn’t be further from Endless Dungeon if you tried. However, this shared sci-fi system plays host to an array of different races and technologies, many of which you’ll encounter in the deep levels of the ship.

Endless Dungeon Review

Endless Dungeon, the latest nugget from development studio Amplitude Studios, is shaping up to be a must-have for video game lovers! Published by publisher SEGA, the game awaits you on October 19, 2023 on PC, PS4, PS5, Xbox One, Xbox Series and Nintendo Switch, marking a new stage in the evolution of tactical roguelikes.

Endless Dungeon: our review of Amplitude Studios' new hit!

Pomerantz LLP is investigating claims on behalf of investors of Amplitude, Inc. ("Amplitude" or the "Company") (NASDAQ:AMPL). Such investors are advised to contact Robert

SHAREHOLDER ALERT: Pomerantz Law Firm Investigates Claims On Behalf of Investors of Amplitude, Inc. - AMPL

LIGO first became world-famous in Sept. 2015, when it detected gravitational waves from merging black holes for the first time. These ripples had traveled for around 1.4 billion years, squashing and squeezing spacetime as they made their way through the universe .

LIGO gravitational wave detector breaks 'quantum limit' to find deep universe black hole collisions

Amplitude

amplitude

It’s 10:15pm. The Amplify conference starts the next morning, and I just found out that I would be attending. I could not wait. I received a pass to Amplify because of the generosity of the Advancing Women in Product (AWIP) community — AWIP is a organization whose goal is to empower the next generation of women product leaders. So there I was on the morning of October 9th, nibbling on my mini breakfast croissant, bumping against the shoulders of product leaders, anticipating the start of a jam-packed, learning-rich, & Ritual Roasters coffee-filled day. If I think about the learnings I drew out of the conference, I’d say they weren’t necessarily profound nuggets of wisdom — they’re more like gentle reminders. Things I already know but tend to forget when building products. Here’s what I learned: Growth is more than a funnel. It’s a lot of different loops. I often break down products via the Pirate Funnel (Awareness, Acquisition, Activation, Retention, Revenue or AAARR for short). But Casey Winters, (VP Product at Eventbrite, formerly growth at Pinterest & Grubhub, among others), argues that that framework prioritizes feeding only into the top of the funnel. When in fact, our products are ongoing feedback mechanisms, aka Loops. Casey Winters interviewing Elena Verna Macro loops stitch our micro loops over time. We have 3 general types of loops: Viral loops are when users organically share or are incentivized to refer other friends back to your product. Content loops are classified based on the of content generated. These can be company (HubSpot, WebMD), user (Pinterest) or supply (Grubhub) generated. type Paid loops vary based on who does the work to distribute the product. These can be ads, sales teams or integrations via partnerships. These growth loops paint a better picture than a funnel, and when architected like this you’ll reap the best outcomes. We can make a difference when we look at products from the lens of social science. I love Dan Ariely (best-selling author of Predictably Irrational, behavioral economics professor at Duke University). I love behavioral economics. I love Dan Ariely’s talk on behavioral economics for products. Dan Ariely giving the keynote! Dan performed a research study in Kenya with the goal of finding ways to help Kenyans save money. He tried several tactics, with mediocre results: injecting friction to make it more difficult to take out money, sending guilt-filled reminders to save, and introducing loss aversion in saving. But the tactic that worked the best by far was so simple and effective: if they saved money that day then they could scratch that day off on a calendar using a coin. The coin method made them aware of their daily actions of saving, and also injected social accountability from their family members. A simple, yet effective change in habit that made the often invisible act of saving money, extremely visible everyday. Another great talk on behavior was from Tatyana Mamut. Her talk offered tactical advice in dealing with bad Internet behavior such as malicious accounts and fake news. As a company scales, entire departments are dedicated to taking down these bad actors before they infect the rest of the user base. These examples were simple yet powerful examples telling us, “hey, our users are humans, too.” We’ve got a lot to learn from a behavioral scientist’s lens. Data’s important, but use it in context. Quantitative data’s so hot these days. And why not? It gives great insights into WHAT our users are doing. But the WHY is also equally important — and we need to marry both quantitative and qualitative data. When initially measuring the success of a product, qualitative data might weigh more heavily. Sarah Tavel from Bessemer ventures emphasizes drawing from our First Principles and front loading user research. Our empathy skills in asking the right questions & observing raw human behavior comes first, until we can really couple the user research with data at scale. “Product managers are the owners of a unique definition of a problem worth solving.” I this definition of a PM from Eric Vishria (General Partner at Benchmark Capital). Let’s break that down. love This definition is great. We can read it and say “oh, yeah, of course.” But we forget to live its principle. We forget to take ownership when things don’t go as planned. We forget to clearly define our goals & problems for everyone else to act on. We forget we are not the solution creators but the conduit to unlock everyone else’s creative potentials. If we can remember this holistic definition, we become better product managers. Some friends from the conference The product leader can understand the systematic complexities of the products they build. I like the analogy that the partners at Benchmark Capital highlighted: We should not lose the forest for the trees. It’s easy to silo into that one northstar metric, or be hyperfocused in building out that one feature. But the best product leaders can look at the forest the complexities of what they’re building. The PM needs to ask themself, PMs need to understand the wider, not often thought of implications of the products they put out into the world. and Does it affect other product metrics? Will this negatively impact specific cultures? What are the ethical risks associated with this design change? As a bonus, my 6th takeaway is that Ritual Coffee Roasters is . (I had one too many cups.) the best Thank you, Advancing Women in Product for the opportunity to learn from all these amazing product minds. I’m already on the waitlist for next years Amplify conference. Congratulations to Amplitude for a successful conference!

Amplify Conference 2018

Authors:
(1) Dorian W. P. Amaral, Department of Physics and Astronomy, Rice University and These authors contributed approximately equally to this work;
(2) Mudit Jain, Department of Physics and Astronomy, Rice University, Theoretical Particle Physics and Cosmology, King’s College London and These authors contributed approximately equally to this work;
(3) Mustafa A. Amin, Department of Physics and Astronomy, Rice University;
(4) Christopher Tunnell, Department of Physics and Astronomy, Rice University. Table of Links Abstract and 1 Introduction 2 Calculating the Stochastic Wave Vector Dark Matter Signal 2.1 The Dark Photon Field 2.2 The Detector Signal 3 Statistical Analysis and 3.1 Signal Likelihood 3.2 Projected Exclusions 4 Application to Accelerometer Studies 4.1 Recasting Generalised Limits onto B − L Dark Matter 5 Future Directions 6 Conclusions, Acknowledgments, and References A Equipartition between Longitudinal and Transverse Modes B Derivation of Marginal Likelihood with Stochastic Field Amplitude C Covariance Matrix D The Case of the Gradient of a Scalar B Derivation of Marginal Likelihood with Stochastic Field Amplitude The full signal in time space is given by Using the series representation of the Bessel function, together with Gamma function identities, the 5 random variables can be integrated out analytically. We arrive at the following marginalized (and normalized) likelihood: where This likelihood can be split into three individual likelihoods for the sum/difference peaks and the Compton peak, as given in Eq. (3.2). The form of the likelihoods for the sum and difference peaks is equivalent. This paper is available on arxiv under CC BY 4.0 DEED license. Authors: (1) Dorian W. P. Amaral, Department of Physics and Astronomy, Rice University and These authors contributed approximately equally to this work; (2) Mudit Jain, Department of Physics and Astronomy, Rice University, Theoretical Particle Physics and Cosmology, King’s College London and These authors contributed approximately equally to this work; (3) Mustafa A. Amin, Department of Physics and Astronomy, Rice University; (4) Christopher Tunnell, Department of Physics and Astronomy, Rice University. Authors: Authors: (1) Dorian W. P. Amaral, Department of Physics and Astronomy, Rice University and These authors contributed approximately equally to this work; (2) Mudit Jain, Department of Physics and Astronomy, Rice University, Theoretical Particle Physics and Cosmology, King’s College London and These authors contributed approximately equally to this work; (3) Mustafa A. Amin, Department of Physics and Astronomy, Rice University; (4) Christopher Tunnell, Department of Physics and Astronomy, Rice University. Table of Links Abstract and 1 Introduction Abstract and 1 Introduction 2 Calculating the Stochastic Wave Vector Dark Matter Signal 2.1 The Dark Photon Field 2.1 The Dark Photon Field 2.2 The Detector Signal 2.2 The Detector Signal 3 Statistical Analysis and 3.1 Signal Likelihood 3 Statistical Analysis and 3.1 Signal Likelihood 3.2 Projected Exclusions 3.2 Projected Exclusions 4 Application to Accelerometer Studies 4 Application to Accelerometer Studies 4.1 Recasting Generalised Limits onto B − L Dark Matter 4.1 Recasting Generalised Limits onto B − L Dark Matter 5 Future Directions 5 Future Directions 6 Conclusions, Acknowledgments, and References 6 Conclusions, Acknowledgments, and References A Equipartition between Longitudinal and Transverse Modes A Equipartition between Longitudinal and Transverse Modes B Derivation of Marginal Likelihood with Stochastic Field Amplitude B Derivation of Marginal Likelihood with Stochastic Field Amplitude C Covariance Matrix C Covariance Matrix D The Case of the Gradient of a Scalar D The Case of the Gradient of a Scalar B Derivation of Marginal Likelihood with Stochastic Field Amplitude The full signal in time space is given by Using the series representation of the Bessel function, together with Gamma function identities, the 5 random variables can be integrated out analytically. We arrive at the following marginalized (and normalized) likelihood: where This likelihood can be split into three individual likelihoods for the sum/difference peaks and the Compton peak, as given in Eq. (3.2). The form of the likelihoods for the sum and difference peaks is equivalent. This paper is available on arxiv under CC BY 4.0 DEED license. This paper is available on arxiv under CC BY 4.0 DEED license. available on arxiv

Derivation of Marginal Likelihood with Stochastic Field Amplitude

Hallo my gals and boys!! Let’s talk alignment. In fast-scaling tech companies, misalignment between product, sales, and marketing isn’t just inconvenient — it’s expensive. Teams operate on different metrics, tools, and timelines. The result? Disconnected strategies, delayed feedback loops, and missed revenue opportunities. The result? Disconnected strategies, delayed feedback loops, and missed revenue opportunities. But alignment isn’t about more meetings — it’s about building the right data infrastructure. building the right data infrastructure In this article, I’ll walk you through how we approached cross-functional GTM alignment by connecting product analytics, marketing attribution, and sales behavior into one growth engine — with practical examples of tools, architecture, and decision-making that drove measurable business impact. Start With the Right Questions (and Data Sources) We began by identifying the key friction points between teams: We began by identifying the key friction points between teams: Marketing wanted to know which campaigns led to product-qualified leads (PQLs), not just MQLs.
Sales needed context: “What features has this user tried? Are they stuck in onboarding?”
Product wanted visibility into which features actually influenced deal velocity or expansion. Marketing wanted to know which campaigns led to product-qualified leads (PQLs), not just MQLs. Sales needed context: “What features has this user tried? Are they stuck in onboarding?” Product wanted visibility into which features actually influenced deal velocity or expansion. To solve this, we mapped 3 core data sources: To solve this, we mapped 3 core data sources: Amplitude for product analytics
HubSpot CRM for sales + lead data
Segment + Google Analytics for marketing attribution Amplitude for product analytics Amplitude HubSpot CRM for sales + lead data HubSpot CRM Segment + Google Analytics for marketing attribution Segment + Google Analytics But plugging in these tools wasn’t enough. The real challenge was connecting them meaningfully. connecting them meaningfully. Build the Unified Growth Data Layer We introduced Segment CDP as the core customer data platform. Every event — from a button click in the product to a LinkedIn ad interaction — was funneled into Segment and tagged with a unified user ID across platforms. Then we pushed this data downstream: Into Amplitude to enrich product dashboards with marketing source + sales status
Into HubSpot to give sales reps real-time visibility into user behavior (“Trial user added 3 teammates yesterday”)
Into a custom Metabase dashboard, which blended marketing source → product usage → sales outcome Into Amplitude to enrich product dashboards with marketing source + sales status Into HubSpot to give sales reps real-time visibility into user behavior (“Trial user added 3 teammates yesterday”) Into a custom Metabase dashboard, which blended marketing source → product usage → sales outcome This gave all teams a shared source of truth: This gave all teams a shared source of truth: → Marketing could optimize campaigns toward feature activations, not just form fills → Sales could prioritize outreach based on product intent → Product could see what GTM levers accelerated usage Automate the Feedback Loops Once this infrastructure was live, we automated the key workflows that previously relied on manual handovers or Slack messages. For example:

When a trial user reached a key milestone (e.g., activated core feature), a HubSpot workflow alerted their assigned rep.
When a high-spend customer downgraded usage, Amplitude triggered a Slack alert to the Customer Success Manager with context.
When a lead converted to a customer, the original UTM parameters and campaign IDs were logged automatically in our analytics schema for revenue attribution. For example: When a trial user reached a key milestone (e.g., activated core feature), a HubSpot workflow alerted their assigned rep.
When a high-spend customer downgraded usage, Amplitude triggered a Slack alert to the Customer Success Manager with context.
When a lead converted to a customer, the original UTM parameters and campaign IDs were logged automatically in our analytics schema for revenue attribution. When a trial user reached a key milestone (e.g., activated core feature), a HubSpot workflow alerted their assigned rep. When a high-spend customer downgraded usage, Amplitude triggered a Slack alert to the Customer Success Manager with context. When a lead converted to a customer, the original UTM parameters and campaign IDs were logged automatically in our analytics schema for revenue attribution. This reduced lag time in handoffs by over 70% and directly contributed to a 22% increase in conversion from trial to paid over three quarters. Make It Actionable, Not Just Visual Dashboards alone aren’t enough. One of our key learnings was: Data must drive action. So we trained each team on how to actually use these insights. Marketing reviewed campaign-driven feature usage weekly, not just CPL. Sales used product activity as part of their deal scoring logic. Product started prioritizing onboarding tweaks based on drop-offs visible in multi-touch funnel analysis. We also created "Growth Syncs" — biweekly sessions with one rep from each GTM team to review data anomalies, test hypotheses, and launch small experiments based on shared insights. "Growth Syncs" — biweekly sessions with one rep from each GTM team to review data anomalies, test hypotheses, and launch small experiments based on shared insights. Conclusion True GTM alignment doesn’t happen in a slide deck — it’s built in your infrastructure. When product, sales, and marketing share a connected data foundation, they stop competing for credit and start collaborating toward growth. This approach helped us scale from siloed departments to a cohesive revenue engine, grounded in real-time insight and proactive decision-making. For anyone building or refining a GTM strategy in a digital company, don’t ask how your teams can align. Ask: What data do they need to think like one team?

Your Marketing Team Doesn’t Need Another Meeting—It Needs a Better Data Stack

Authors:
(1) Anh Pham, Deloitte Consulting LLP;
(2) Andrew Vlasic, Deloitte Consulting LLP. Authors: Authors: (1) Anh Pham, Deloitte Consulting LLP; (2) Andrew Vlasic, Deloitte Consulting LLP. Table of Links Abstract and I. Introduction Abstract and I. Introduction II. Overview of error Mitigation Methods II. Overview of error Mitigation Methods III. Methodology III. Methodology IV. Results and Discussions, and References IV. Results and Discussions, and References The study examines the efficacy of the Fire Opal error suppression and AI circuit optimization system integrated with IBM’s quantum computing platform for a multi-modal distribution loading algorithm. Using Kullback-Leibler (KL) divergence as a quantitative error analysis, the results indicate that Fire Opal can improve on the time-dependent distributions generated by our Conditional Quantum Generative Adversarial algorithm by 30-40% in comparison with the results on the simulator. In addition, Fire Opal’s performance remains consistent for complex circuits despite the needs to run more trials. The research concludes that Fire Opal’s error suppression and circuit optimization significantly enhanced quantum computing processes, highlighting its potential for practical applications. In addition, the study also reviews leading error mitigation strategies, including zero noise extrapolation (ZNE), probabilistic error cancellation (PEC), Pauli twirling, measurement error mitigation, and machine learning methods, assessing their advantages and disadvantages in terms of technical implementation, quantum resources, and scalability. I. INTRODUCTION Data loading is critical for many quantum algorithms and applications. However, it is a challenging problem when executing on NISQ quantum hardware due to the longer circuit depth in this quantum subroutine. In this report, we demonstrate the utility of error suppression and AI-enabled circuit optimization in Fire Opal to load multi-modal distributions on IBM Kyoto as implemented in our Conditional Quantum Generative Adversarial Network (C-QGAN) algorithm. Specifically, the distributions generated with Fire Opal on quantum hardware produced results much closer to the ideal distributions, and showed an approximate 30% − 40% improvement in results relative to running without it based KL divergence analysis. Previously we proposed a new quantum algorithm known as C-QGAN [1] to load multiple non-uniform distributions using condition registers. This data loading technique was then applied in conjunction with quantum amplitude estimation [2] to evaluate complex financial instrument called Asian option. The state preparation based on C-QGAN was shown to be less computationally expensive than other well known technique like GroverRudolph [3] and QGAN [4] which can potentially negate the quadratic speedup of algorithm like QAE [5] when they are approximating stochastic processes. Due to the noisy nature of current NISQ hardware, various error mitigation techniques have been applied to enable improve the outputs when quantum algorithms are run on quantum hardware. Error mitigation techniques are post-processing algorithms, which happen at the software level, that can improve on the distorted values obtained from quantum hardware due to different noise sources. In addition, these techniques have been applied to improve in many variational quantum algorithms [6] which have many important applications in quantum machine learning, chemistry and optimization. However, a potential drawbacks of many error mitigation techniques can be that they require extra classical and quantum computational resources to recover the noisefree values, which limits their applications only for shortdepth circuits [7]. As a result, our report aims to explore error suppression technique, which happens at the hardware level, as implemented in Fire Opal to understand their utility in the state-preparation process. This paper is available on arxiv under CC by 4.0 Deed (Attribution 4.0 International) license. This paper is available on arxiv under CC by 4.0 Deed (Attribution 4.0 International) license. available on arxiv available on arxiv

How Fire Opal Makes Quantum Data Loading Smarter and More Accurate

Authors:
(1) VINÍCIUS YU OKUBO, Dept. Electronic Systems Engineering, Polytechnic School, University of São Paulo, Brazil;
(2) KOTARO SHIMIZU, Department of Applied Physics, The University of Tokyo, Tokyo 113-8656, Japan;
(3) B. S. SHIVARAM, Department of Physics, University of Virginia, Charlottesville, Virginia 22904, USA;
(4) HAE YONG KIM, Dept. Electronic Systems Engineering, Polytechnic School, University of São Paulo, Brazil. Authors: Authors: (1) VINÍCIUS YU OKUBO, Dept. Electronic Systems Engineering, Polytechnic School, University of São Paulo, Brazil; (2) KOTARO SHIMIZU, Department of Applied Physics, The University of Tokyo, Tokyo 113-8656, Japan; (3) B. S. SHIVARAM, Department of Physics, University of Virginia, Charlottesville, Virginia 22904, USA; (4) HAE YONG KIM, Dept. Electronic Systems Engineering, Polytechnic School, University of São Paulo, Brazil. Table of Links Abstract and I. Introduction Abstract and I. Introduction II. Related Work II. Related Work III. Methodology III. Methodology IV. Experiments and Results IV. Experiments and Results V. Conclusion and References V. Conclusion and References II. RELATED WORKS A. JUNCTIONS AND TERMINALS DETECTION A. JUNCTIONS AND TERMINALS DETECTION A. JUNCTIONS AND TERMINALS DETECTION Junctions and terminals are shapes with relevance that extend beyond materials. Within computer vision, recognizing and enumerating them has been performed in diverse contexts, such as natural landscapes [20], biology [21] and handwriting images [22]. B. CLASSICAL METHODS FOR JUNCTIONS AND TERMINALS DETECTION B. CLASSICAL METHODS FOR JUNCTIONS AND TERMINALS DETECTION B. CLASSICAL METHODS FOR JUNCTIONS AND TERMINALS DETECTION One prevalent approach for detecting junctions and terminals involves using skeletonization as a pre-processing step. This process reduces the image to one-pixel-wide lines to represent its structures. Points in skeleton can then be identified as terminals, junctions and crossings based on their neighboring pixels. This technique has been applied in vascular images [23] and in the analysis of handwritten Chinese characters [24]. Pre-processing techniques using contour information have also been explored for junction detection. Lee and Wu [25] investigated stroke extraction in Chinese characters. Their method segments regions according to their contour, identifying junctions by counting neighboring regions. Maire et al. [26] proposed a junction detector in natural images by locating intersecting contours. This approach applies an expectation–maximization style algorithm to iteratively select relevant contours and suggest the junction’s position. However, skeletonization and contour finding are noise sensitive processes, and a pre-processing error will lead to a detection error. Junctions and terminals can also be identified by analyzing the arrangement of linear structures within the image. Su et al. [27] describe a technique for identifying these linear structures using the Hessian Matrix. This approach has been validated in biological images such as blood vessels, neutrites and tree branches. Xia et al. [20] present a junction detection method in natural images, based on amplitudes and phases of the normalized gradients of the image. Template-based approaches quantify the similarity of the appearance of image regions and the template. Deriche and Blaszka [28] modeled this approach as energy minimization, which is calculated by the deviation between the image and a predetermined model. This enabled the detection of key image features, such as edges, corners and terminals. C. DEEP LEARNING METHODS FOR JUNCTIONS AND TERMINALS DETECTION C. DEEP LEARNING METHODS FOR JUNCTIONS AND TERMINALS DETECTION C. DEEP LEARNING METHODS FOR JUNCTIONS AND TERMINALS DETECTION Owing to the success of R-CNN based detection techniques, Pratt et al. [21] developed a pipeline for identifying junctions and crossings in retinal vascular structures. Their method involves two main steps: initially, detection regions are proposed centred along the blood vessels, which are then classified as junctions, crossings or background. To generate the detection regions, their approach requires a binary segmented version of the exam. Theses images undego a skeletonization process, with the resulting points serving as references for the centers of the blood vessels. Zhao et al. [29], addressing the same problem, proposed using a Mask R-CNN base model [30] for region proposal. This strategy enables inference without the need for binary segmented version of the exam. However, during training, the segmented images are still used in the Mask R-CNN model to enhance its learning capabilities. This approach surpassed the performance of previous techniques in the detection of junctions and crossings in retinal vascular images. III. METHODOLOGY A. DATASET A. DATASET A. DATASET In this study, we used films of a ferromagnetic material of recognized technological importance, Bi:YIG, and obtained magnetic images of the labyrinthine patterns using a microscope with polarized light [31]. We specifically focused on the evolution of the labyrinthine patterns under the demagnetization field protocol described below. First, we prepared the sample in the fully saturated state by applying a sufficiently large magnetic field in +z direction, which is perpendicular to the films. In this state, magnetic moments in the Bi:YIG film are forced to point in the field direction. Next, we instantaneously dropped the field to zero and hold it for 10 seconds to get the image. Since the magnetic field is zero, magnetic moments can point upward and downward, resulting in the labyrinthine patterns shown in Fig. 1; the bright and dark regions represent the domains with opposite directions of magnetic moments. Such a process involving switching on and off the magnetic fields is considered half of the demagnetization step. In the remaining half step, a magnetic field was applied with a reduced amplitude and oriented in the opposite direction. We again captured the magnetic domain image after reducing the magnetic field to zero. By repeating these protocols up to 18 steps, we investigated the evolution of the labyrinthine patterns in the demagnetization process step by step. The amplitude of the magnetic field was exponentially reduced with each step. We conducted a series of demagnetization processes from the fully saturated state, repeating this cycle six times. Furthermore, we explored another situation, where the magnetic field was initially applied in the −z direction, and its direction was alternated step by step. Consequently, a total of 12 demagnetization processes were performed, yielding a collection of 444 domain images. All measurements reported here were performed at room temperature. The experimentally obtained images covered an area of 2 mm × 1.8 mm. The original high-resolution color images were converted to grayscale and their resolutions were reduced to 1300×972 to facilitate processing. Furthermore, a median filter with kernel size 3 was applied to reduce noise. B. TM-CNN OVERVIEW B. TM-CNN OVERVIEW B. TM-CNN OVERVIEW Our approach to detect junctions and terminals in magnetic labyrinthine patterns consists of two sequential steps: proposal of potential detections and their classification between junction, terminal and false detections. It is inspired by other cascaded object detection techniques like Viola and Jones face detection [8], R-CNN [12], and scale and rotation invariant template matching [32]. Fig. 3 illustrates the overall structure of TM-CNN. In the first phase, the algorithm generates a preliminary set of potential detections. It must propose all true defects, even if it also generates many false positives. We achieve this by applying template matching detection with a low threshold, followed by a non-maximum suppression. In the second phase, in order to eliminate the false positives, each potential detection is filtered by a CNN classifier. C. TEMPLATE MATCHING C. TEMPLATE MATCHING C. TEMPLATE MATCHING The basic form of template matching finds instances of a smaller template T within a larger image I. This is done by calculating some similarity metric between the model T and the content of a moving window located at each possible position of I. We measured the similarity between the image and the template using the Normalized Cross Correlation (NCC). NCC is invariant to linear changes in brightness and/or contrast. NCC between template T and image I at pixel (x, y) is calculated as: This basic approach is not well suited for detecting junctions and terminals in the magnetic labyrinthine structures because a single template is not capable of modeling: All possible rotations;


All deformed shapes of defects. All possible rotations; All possible rotations; All deformed shapes of defects. All deformed shapes of defects. To solve problem (1), we employ a rotation-invariant template matching based on exhaustive evaluation of rotated templates. There are some alternative rotation-invariant techniques based on circular and radial projections [32], [34], and on Fourier coefficients of circular and radial projections [35]. These techniques can reduce computational requirements, but their implementations are complex and require parameter tuning. Furthermore, our application does not require exceptional computational performance, as the processing is offline. Thus, our technique uses the standard OpenCV[1] template matching implementation, which is highly optimized using FFT and special processor instructions. D. TEMPLATES AND MASKS USED IN THE EXPERIMENT D. TEMPLATES AND MASKS USED IN THE EXPERIMENT D. TEMPLATES AND MASKS USED IN THE EXPERIMENT We manually designed templates and masks, and empirically tuned them to capture various junction and terminal shapes. They have 21×21 resolutions and are generated at runtime. Templates represent magnetic strips as black lines radiating from their center drawn on a white background. Meanwhile, masks are created from black backgrounds with white areas defining relevant coordinates used in template matching. Their main purpose is to obscure the space between strips and background, addressing variations in widths and curvatures, and to limit the background regions to reduce interference from neighboring strips. Table 1 exemplifies the templates and masks used to detect junctions and terminals. Template matching is applied separately for each (template, mask) pair. Several template matchings are computed in parallel using the OpenMP library[2]. This process takes about 80 seconds to process an image on an i7-9750H processor. To obtain the final correlation map corr, we calculate the maximum value among all n = 3 × 439 + 5 × 120 = 1917 NCC maps for each position (x, y), that is: Pixels where corr(x, y) exceeds a predefined threshold t are considered potential detections. However, a single junction/terminal may encompass multiple neighboring points with correlation values greater than the threshold t. Therefore, it is necessary to perform some form of non-maximum suppression to eliminate duplicate detections and select only the true center of junction/terminal. Kim et al. [18] present a solution to this problem: Whenever two potential detection points p1 and p2 are separated by a distance smaller than a threshold c, the point with the lowest correlation value is discarded. In this work, we use a slightly different approach, but with the same practical result: Whenever the algorithm finds a potential detection, it executes a breadth-first search algorithm. This algorithm recursively searches adjacent pixels (x, y) where the correlation exceeds 80% of the threshold (that is, corr(x, y) > 0.8t) and saves the pixel with the highest correlation. Subsequently, the searched area has its correlation value set to zero to avoid re-detection. Fig. 3d highlights the searched area in red. The pixel with the highest correlation is chosen as the center of the junction/termination. This process performs detection in a single pass. E. DATASET ANNOTATION E. DATASET ANNOTATION E. DATASET ANNOTATION To classify potential detections into true or false using a CNN classifier, we must first create annotated training images. TM-CNN makes it easier to create training examples, as it allows one to annotate the examples semi-automatically. This process is divided in two phases. 1) Template matching-assisted annotation 1) Template matching-assisted annotation In this phase, only a small set of images are annotated. Initially, we apply template matching followed by nonmaximum suppression to identify the centers of possible detections, together with their probable labels (junction or terminal). Without this help from template matching, we would have to manually and precisely locate the centers of thousands of defects. Subsequently, a human reviewer makes corrections to ensure that the labels given by the template matching are correct. The reviewer may change the labels to junction, terminal or false detection. After all positive detections are annotated along with a small set of false detections, a larger set of false detections is created by lowering the template matching threshold and sampling new false detections. These images, now with positive and negative annotations, are used to train a preliminary version of the CNN classifier. 2) Deep learning-assisted annotation 2) Deep learning-assisted annotation Due to the small number of images in the initial training set, the preliminary CNN classifier cannot accurately classify all magnetic stripe defects. Nonetheless, this preliminary model is integrated into the annotation procedure to alleviate the required workload. In the second phase, we continue using template matching to generate the initial set of detections. However, the preliminary CNN classifier is employed to identify most of the template matching errors, thus speeding up the annotation process. As new images are annotated, more accurate models are trained to further simplify the annotation workload. The final training set consists of 17 images derived from a single annealing protocol, selected to cover varied experimental configurations of ascending and descending magnetic fields at different magnitudes. Out of these, 16 were selected from the quenched (unordered) state, as they represent a more diverse set of shapes and represent a greater challenge for classification. The training set encompasses a total of 33,772 detections, which includes 12,144 junctions, 12,777 terminals, and 8,851 false detections. F. CANDIDATE FILTERING BY CNN F. CANDIDATE FILTERING BY CNN F. CANDIDATE FILTERING BY CNN Our algorithm extracts small 50×50 patches centered around each detection point and a CNN classifies them into three classes: junction, terminal or false positive. The size of patches for CNN classification is larger than the size of template matching models (21×21), allowing CNN to use more contextual information than template matching. We use a simple CNN model to classify the small patches (Fig. 4). It has four convolutional layers with 32, 64, 128 and 256 filters, all using 3×3 kernels. The first three convolutional layers are followed by max pooling layers to downsample the feature maps and global max pooling is applied after the last convolutional layer. This is followed by dropout and two fully connected layers: the first with 128 nodes and the second with three output nodes. The ReLU activation function is used across the model, except at the output layer where the softmax is used. In total, this network has only 422,608 parameters. For comparison, VGG-16 and ResNet-50 (common backbones for detection) have 138 million and 23 million parameters, respectively. Thanks to its simplicity, our model is fast and can make predictions even without GPUs and takes around 30 seconds for filtering each image using an i7-9750H processor with 16 GB of RAM. This paper is available on arxiv under CC BY 4.0 DEED license. This paper is available on arxiv under CC BY 4.0 DEED license. available on arxiv [1] Open Source Computer Vision Library, https://opencv.org [2] Open Multi-Processing, https://www.openmp.org/ [3] https://www.tensorflow.org/

Why TM-CNN Is a Material Scientist's New Best Friend

Authors:
(1) Smaran Deshmukh, Email: smarandeshmukh@gmail.com;
(2) Jayant Murthy, Indian Institute of Astrophysics, Bangalore 506034, India (murthy@iiap.res.in). Authors: Authors: (1) Smaran Deshmukh, Email: smarandeshmukh@gmail.com; (2) Jayant Murthy, Indian Institute of Astrophysics, Bangalore 506034, India (murthy@iiap.res.in). Table of Links Abstract and 1 Introduction Abstract and 1 Introduction 2. Summary of the Story 2. Summary of the Story 3. Observables 3. Observables 4. Possible Scenarios and Consequences 4. Possible Scenarios and Consequences The other Suns The other Suns Simulation Simulation Conclusion and References Conclusion and References SIMULATION We have written a program computer program that simulates a stellar system. We have not attempted to match the scale of the system, this is only a demonstration. The user inputs the initial positions and velocities of the stars, as well as a precision parameter which determines how often forces and velocities of the individual bodies are updated and recalculated. The simulation is based on Newton’s laws of motion: We have not attempted to match the scale of the system, this is only a demonstration where ‘r’ and ‘r0’ are the instantaneous and initial positions, ‘v’ and ‘v0’ are the instantaneous and initial velocities, ‘a’ is the instantaneous acceleration and ‘dt’[1] is the time step in seconds. The smaller the value of dt, the better the simulation but the longer the simulation length. KALGASH 2 DISTANCE We will suppose that Kalgash 2 is a gaseous moon slightly less dense than Saturn (0.687g/cm3). Since it has the same mass as the earth, it will have a diameter of 26,540 km. It also subtends an angle of 26”. This means that Kalgash 2 is at a distance of 3.04 million kilometres. However, both the planet and the moon are light compared to the stars and we can ignore them in the simulation. As discussed above, the stellar system of Asimov’s Nightfall has 6 stars but with two sets of binaries (Trey-Patru and Tano-Sitha), which we treat as single point masses. Although different stable configurations may be possible, we have worked out one possible configuration[2] which we found to be stable for a few hundred years. We further arrange these stars into three sets of 2-star systems. First, Onos and Dovim (circling each other) as one independent system, then Tano-Sitha and Onos-Dovim (Figure 5), where the forces act on the centre of masses of the two binaries. And finally, Trey-Patru and Onos-Dovim-Tano-Sitha. The graph is similar to a velocity graph of a binary star (in this case Onos-Dovim is the binary, since we have treated TS as a single star). The amplitudes of the oscillation slowly increase, but this is within the limits of the precision settings of our program. Also the amplitudes increase over a period of several hundred years: the people of Kalgash need have no fear of being swallowed by a sun! Next, we add Trey-Patru. Now we suspect that the addition of Trey-Patru will quickly rip the system. To counter this, we make sure that Trey-Patru is not close to Onos-Dovim for large periods of time. This is done by giving TP an opposite angular velocity as the following image shows: The addition of TP increases the tidal forces on Onos-Dovim. However, for the duration of the story (about 2 years) these effects will not be perceptible. (Though subsequent generations of the Kalgash people will face dangerous scenarios.) The Kalgash people measured the universe to be 110 light minutes across. Perhaps, they made this measurement at a very specific time in the evolution of the system. Luckily for them, the hottest binary star maintains a safe distance from them. The large amplitudes mean that at some later time, the “stars” might be become visible. CONCLUSION We have explored several aspects of Asimov’s novel. We have found that the suns, especially Dovim are bright enough to blot out the stars. Kalgash 2 can eclipse Dovim for a period of 9 hours. We also tested one possible star configuration and after running some simulations, we found that the system is possible for short periods of time. Several other configurations might exist and have to be explored more fully. REFERENCES Asimov, I., 1941. Nightfall. [Online] Available at: https://www.uni.edu/morgans/astro/course/nightfall. pdf [Accessed April 2014]. Asimov, I. & Silverberg, R., 1991. Nightfall. New York, NY: Bantam Books. Jens Buus, J. M., n.d. Latitude Dependence of the Maximum Duration of a total Solar Eclipse. [Online] Available at: www.eclipsechasers.com/papers/Maximum_duration.pdf [Accessed 11 July 2014]. Meeus, J., 2003. The Maximum Possible Duration of a Total Solar Eclipse. Journal of the British Astronomical Association, 113(6), pp. 343-348. Orosz, J. A. et al., 2012. Sciencemag. [Online] Available at: https://www.sciencemag.org/content/337/6101/1511. short [Accessed May 2014]. Roatsch, T., Jaumann, R. & Thomas, P., 2009. Wikipedia - Cartographic Mapping of the Icy Satellites Using ISS and VIMS Data. [Online] Available at: http://en.wikipedia.org/wiki/Tethys_%28moon%29 [Accessed 11 July 2014]. Sagan, C., 1992. Nature. [Online] Available at: http://www.nature.com/nature/journal/v357/n6374/ pdf/357113a0.pdf [Accessed May 2014]. Wikipedia, 2014. Apparent Magnitude. [Online] Available at: http://en.wikipedia.org/wiki/Apparent_magnitude [Accessed July 2014]. Williams, D. D. R., 2004. Wikipedia. [Online] Available at: http://en.wikipedia.org/wiki/Apparent_magnitude [Accessed May 2014]. Williams, D. D. R., 2014. Nasa. [Online] Available at: http://nssdc.gsfc.nasa.gov/planetary/factsheet/index.h tml [Accessed 11 July 2014]. Williams, D. R., 2006. Wikipedia. [Online] Available at: http://en.wikipedia.org/wiki/Saturn [Accessed 11 July 2014]. This paper is available on arxiv under CC BY 3.0 DEED license. This paper is available on arxiv under CC BY 3.0 DEED license. available on arxiv [1] The value of dt is critical in the simulations. We have used 1000sec. This means that the errors will become significant over large periods of time. The simulations here just give an approximate idea of the system stability. [2] In this configuration we have changed some distances to allow a stable configuration.

Can Kalgash Exist? Simulating the Stellar System of Asimov’s Sci-Fi Classic

I Built a €500K Neobank for Immigrants—It Didn't Go As Planned

Ek het 'n €500K Neobank vir immigrante gebou—dit het nie gegaan soos beplan nie

Създадох Neobank за 500 000 евро за имигранти — не мина по план

Creé un neobanco de 500.000 € para inmigrantes y no salió como estaba previsto

Építettem egy 500 000 eurós Neobankot a bevándorlók számára – nem úgy ment, ahogy terveztem

Saya Membangun Neobank Senilai €500K untuk Imigran—Tidak Berjalan Sesuai Rencana

移民向けに50万ユーロのネオバンクを構築したが、計画通りにはいかなかった

ຂ້ອຍສ້າງ Neobank 500,000 ເອີໂຣສຳລັບຄົນເຂົ້າເມືອງ—ມັນບໍ່ເປັນໄປຕາມແຜນການ.

Eu construí um Neobanco de € 500 mil para imigrantes — não saiu como planejado

මම සංක්‍රමණිකයින් සඳහා යුරෝ 500,000 ක නියෝබැංකුවක් හැදුවා - එය සැලසුම් කළ පරිදි සිදු වුණේ නැහැ.

Ndakavaka € 500K Neobank yeVatorwa-Haina Kuenda Sezvatakarongwa

Jag byggde en €500 000 Neobank för invandrare – det gick inte som planerat

Ngakhe i-Neobank engu-€500K yabafuduki—Ayihambanga Njengoba Bengihleliwe

Can We Simulate Quantum Phenomena? The FPGA Coprocessor Experiment

Table of Links Abstract and 1. Introduction Clock in Orbit
2.1 Coordinate Time
2.2 Local Frame for the Moon


Clock Rate Differences Between Earth and Moon


Clocks at Earth-Moon Lagrance Points
4.1 Clock at Lagrange point L1
4.2. Clock at Lagrange point L2
4.3. Clock at Lagrange point L4 or L5


Conclusions Appendix 1: Fermi Coordinates with Origin at the Center of the Moon Appendix 2: Construction of Freely Falling Center of Mass Frame Appendix 3: Equations of Motion of Earth and Moon Appendix 4: Comparing Results in Rotating and Non-Rotating Coordinate Systems Acknowledgments and References 3. CLOCK RATE DIFFERENCES BETWEEN EARTH AND MOON The Earth and the Moon orbit around their mutual center of mass in different Keplerian orbits. Meanwhile, the center of mass of the Earth-Moon system orbits around the Sun in an approximately Keplerian orbit. To calculate the rate differences between clocks on Earth and on the Moon, a fictitious locally freely-falling inertial frame is introduced at the Earth-Moon center of mass. This makes it convenient to calculate the proper times elapsed on moving clocks in terms of Keplerian motions of the Earth and the Moon. The Sun’s contribution is only tidal effects. If we omit the tidal potential of the Sun, the scalar invariant takes a simple form (Appendix 2) Consider a clock fixed on the surface of the rotating geoid of Earth. Since the geoid is a surface of approximate hydrostatic equilibrium, if such clocks are viewed from the local inertial frame they beat at the same rate, which can be evaluated at the equator. The proper time on the Earth-based clock becomes The first two contributions can be identified with the quantity Φ0. The contribution from the Moon’s potential can be well approximated by setting where D is the Earth-Moon distance. The dot product term between velocities will depend on the specific position of the clock and will vary with a daily period; this variation is similar to the corrections to the gravitational potential contribution from the Moon arising from the fact that the clock is not at the center of the Earth. Omitting such contributions give A similar argument applied to a clock fixed on the rotating Moon’s surface of hydrostatic equilibrium gives the proper time The following combination of quantities occurs frequently and can be reduced to a simpler expression The difference of squares of velocities is then Using Eq.(27) in Eq.(18), we obtain This term will have a period of approximately 27 days. Cross terms between velocities of the center of mass and rotational velocities have been omitted. One such term is This contribution will also vary with the rotation period of the Earth. Another such term is Since the Moon is tidally locked to the Earth, its center of mass velocity and the velocity of a clock on the selenoid will be highly correlated. Therefore, this term might give rise to a constant long-term average. Omitting the position-dependent terms, we have used the constants listed in Table I to evaluate the constant contribution and the amplitude of the periodic term. We find: None of the above estimates include tidal effects. This omission is because as a tidal force pushes back and forth on a satellite, two other side effects have to be accounted for. These are a change in the position of the satellite that entails a change in the gravitational potential of the body about which the satellite is orbiting and a change in the velocity of the satellite clock that changes its second-order Doppler shift. The residuals of the gravitational potential and second-order Doppler shift for the Earth-Moon system obtained by subtracting the Keplerian model from that obtained from DE440 are graphed in Fig. 1. Previous work on such problems has shown that these changes are of similar orders of magnitude. Summarizing, there are three contributions to the frequency shift of a clock in a satellite that are of similar orders of magnitude: (1) the perturbing tidal potential itself; (2) the perturbed position that changes the contribution from the main potential; (3) the perturbed velocity that changes the time dilation contribution. Although the perturbing tidal potential can easily be estimated, calculating the other two contributions is more complicated. When the Keplerian model is compared with DE440 ephemerides, the effects of solar tides are plotted in Fig. 2. It might appear that the second-order Doppler contribution to the rate difference depends on the coordinate system used. In the center-of-mass coordinates, a difference of squares of velocities appears, but in a system in which the Earth is at rest, the square of the relative velocity appears. Appendix 4 shows that contributions from the centrifugal potential, which occurs in a rotating coordinate system resolve the apparent discrepancy. Authors:
(1) Neil Ashby, National Institute of Standards and Technology, Boulder, CO 80305 (Neil.Ashby@Colorado.Edu);
(2) Bijunath R. Patla, National Institute of Standards and Technology, Boulder, CO 80305 (bijunath.patla@nist.gov). This paper is available on arxiv under CC0 1.0 license. Table of Links Abstract and 1. Introduction Abstract and 1. Introduction Clock in Orbit
2.1 Coordinate Time
2.2 Local Frame for the Moon Clock Rate Differences Between Earth and Moon Clocks at Earth-Moon Lagrance Points
4.1 Clock at Lagrange point L1
4.2. Clock at Lagrange point L2
4.3. Clock at Lagrange point L4 or L5 Conclusions Clock in Orbit 2.1 Coordinate Time 2.2 Local Frame for the Moon Clock in Orbit Clock in Orbit 2.1 Coordinate Time 2.1 Coordinate Time 2.2 Local Frame for the Moon 2.2 Local Frame for the Moon Clock Rate Differences Between Earth and Moon Clock Rate Differences Between Earth and Moon Clock Rate Differences Between Earth and Moon Clocks at Earth-Moon Lagrance Points 4.1 Clock at Lagrange point L1 4.2. Clock at Lagrange point L2 4.3. Clock at Lagrange point L4 or L5 Clocks at Earth-Moon Lagrance Points 4.1 Clock at Lagrange point L1 4.1 Clock at Lagrange point L1 4.2. Clock at Lagrange point L2 4.2. Clock at Lagrange point L2 4.3. Clock at Lagrange point L4 or L5 4.3. Clock at Lagrange point L4 or L5 Conclusions Conclusions Conclusions Appendix 1: Fermi Coordinates with Origin at the Center of the Moon Appendix 1: Fermi Coordinates with Origin at the Center of the Moon Appendix 2: Construction of Freely Falling Center of Mass Frame Appendix 2: Construction of Freely Falling Center of Mass Frame Appendix 3: Equations of Motion of Earth and Moon Appendix 3: Equations of Motion of Earth and Moon Appendix 4: Comparing Results in Rotating and Non-Rotating Coordinate Systems Appendix 4: Comparing Results in Rotating and Non-Rotating Coordinate Systems Acknowledgments and References Acknowledgments and References 3. CLOCK RATE DIFFERENCES BETWEEN EARTH AND MOON The Earth and the Moon orbit around their mutual center of mass in different Keplerian orbits. Meanwhile, the center of mass of the Earth-Moon system orbits around the Sun in an approximately Keplerian orbit. To calculate the rate differences between clocks on Earth and on the Moon, a fictitious locally freely-falling inertial frame is introduced at the Earth-Moon center of mass. This makes it convenient to calculate the proper times elapsed on moving clocks in terms of Keplerian motions of the Earth and the Moon. The Sun’s contribution is only tidal effects. If we omit the tidal potential of the Sun, the scalar invariant takes a simple form (Appendix 2) Consider a clock fixed on the surface of the rotating geoid of Earth. Since the geoid is a surface of approximate hydrostatic equilibrium, if such clocks are viewed from the local inertial frame they beat at the same rate, which can be evaluated at the equator. The proper time on the Earth-based clock becomes The first two contributions can be identified with the quantity Φ0. The contribution from the Moon’s potential can be well approximated by setting where D is the Earth-Moon distance. The dot product term between velocities will depend on the specific position of the clock and will vary with a daily period; this variation is similar to the corrections to the gravitational potential contribution from the Moon arising from the fact that the clock is not at the center of the Earth. Omitting such contributions give A similar argument applied to a clock fixed on the rotating Moon’s surface of hydrostatic equilibrium gives the proper time The following combination of quantities occurs frequently and can be reduced to a simpler expression The difference of squares of velocities is then Using Eq.(27) in Eq.(18), we obtain This term will have a period of approximately 27 days. Cross terms between velocities of the center of mass and rotational velocities have been omitted. One such term is This contribution will also vary with the rotation period of the Earth. Another such term is Since the Moon is tidally locked to the Earth, its center of mass velocity and the velocity of a clock on the selenoid will be highly correlated. Therefore, this term might give rise to a constant long-term average. Omitting the position-dependent terms, we have used the constants listed in Table I to evaluate the constant contribution and the amplitude of the periodic term. We find: None of the above estimates include tidal effects. This omission is because as a tidal force pushes back and forth on a satellite, two other side effects have to be accounted for. These are a change in the position of the satellite that entails a change in the gravitational potential of the body about which the satellite is orbiting and a change in the velocity of the satellite clock that changes its second-order Doppler shift. The residuals of the gravitational potential and second-order Doppler shift for the Earth-Moon system obtained by subtracting the Keplerian model from that obtained from DE440 are graphed in Fig. 1. Previous work on such problems has shown that these changes are of similar orders of magnitude. Summarizing, there are three contributions to the frequency shift of a clock in a satellite that are of similar orders of magnitude: (1) the perturbing tidal potential itself; (2) the perturbed position that changes the contribution from the main potential; (3) the perturbed velocity that changes the time dilation contribution. Although the perturbing tidal potential can easily be estimated, calculating the other two contributions is more complicated. When the Keplerian model is compared with DE440 ephemerides, the effects of solar tides are plotted in Fig. 2. It might appear that the second-order Doppler contribution to the rate difference depends on the coordinate system used. In the center-of-mass coordinates, a difference of squares of velocities appears, but in a system in which the Earth is at rest, the square of the relative velocity appears. Appendix 4 shows that contributions from the centrifugal potential, which occurs in a rotating coordinate system resolve the apparent discrepancy. Authors: (1) Neil Ashby, National Institute of Standards and Technology, Boulder, CO 80305 (Neil.Ashby@Colorado.Edu); (2) Bijunath R. Patla, National Institute of Standards and Technology, Boulder, CO 80305 (bijunath.patla@nist.gov). Authors: Authors: (1) Neil Ashby, National Institute of Standards and Technology, Boulder, CO 80305 ( Neil.Ashby@Colorado.Edu ); Neil.Ashby@Colorado.Edu (2) Bijunath R. Patla, National Institute of Standards and Technology, Boulder, CO 80305 ( bijunath.patla@nist.gov ). bijunath.patla@nist.gov This paper is available on arxiv under CC0 1.0 license. This paper is available on arxiv under CC0 1.0 license. available on arxiv available on arxiv

The Role of Gravity and Motion in Space-Time Coordination

You have a lead scoring model in your CRM that just simply is not working? The harsh truth is that your lead scoring system as it stands, is probably broken. You are assigning points for job titles, for the number of times people visit your website, or open your emails, but when your sales team goes to these so-called “hot leads,” a lot of these people are not even interested. Sound familiar? The problem is that traditional lead scoring systems are based on the assumption that buyers go on a simple, linear journey. The reality is that today’s buyers are much more complex. By the time a buyer needs a solution, the decision is often already made. They do their homework long before you can even pronounce their name—glossing over reviews, measuring competitors, exploring industry blogs.


Their engagement is inconsistent — they may view a lot of content one week and none the next.


They may not respond to your emails, but engage with your product often — a clear sign of intent missed by static scoring models. If you're still relying on static lead scoring, you might be missing out on valuable opportunities. This is where Dynamic Lead Scoring comes in - a more innovative, real-time approach that prioritizes leads based on what buyers do and intent signals. This guide will give you some ideas for how to create a dynamic lead-scoring model in your CRM. Then, by using AI, behavioral data, and automation, you can make sure your sales team is spending its time on the leads that are most likely to convert. Ready to get started? Let’s dive in. 🚀 Static and Dynamic Lead Scoring: What’s the Difference? Feature

Static Lead Scoring

Dynamic Lead Scoring



Data Sources

Demographics, firmographics, past actions

Real-time engagement, behavioral patterns, AI insights



Updates

Manual, periodic

Continuous, automated



Intent Signals

Clicks, form fills

Product usage, time-sensitive actions, in-market signals



Sales Alignment

Often misaligned

Predictive and relevant Static models make fixed point assignments (i.e., “10 points for an email open”), whereas dynamic scoring updates in real-time based on actual buyer behavior, providing more relevant and actionable insights. Step 1: Identify Your Dynamic Lead Scoring Model When creating a dynamic scoring model, there are three main types of data that you need to focus on: Fit Score (Who They Are)
IndustryCompany size
Job title
Company Revenue 👉 CRM Setup Tip: If your CRM is missing company data, identify gaps and use enrichment tools to fill them. Next, build workflows to move this data into contacts and create baseline scores and segments against your ICP (Ideal Customer Profile). Engagement Score (How They Interact With You)
Visits to your website (e.g. product or pricing pages = high intent)Content downloads
Webinar attendance
Email interactions
Social media engagement 👉 CRM Setup Strategy: Use a marketing automation tool like HubSpot, Marketo, Pardot, etc., to track engagement, and automatically update scores dynamically. Intent Score (How Ready they are to Buy)
Engagement on the product (especially for SaaS companies)Frequent website visits
Competitor comparisons (activity on G2 or Capterra, etc.)
High engagement with sales materials (case studies, ROI calculators, etc.)
More web searches aligning with your product 👉 CRM Setup Hack: Link intent data solutions, such as HubSpot Breeze, 6sense, Bombora, or others to your CRM. Step 2: Add Scoring Automation in Your CRM Option 1: HubSpot 3rd Party Data Enrichment: Use ZoomInfo, HubSpot Breeze, Bombora, etc. to track buyer intent outside of your own channels.


Establish Scoring Models – Fit, Engagement, and Intent scores should each have their own models.


Sales Alerts: Automatically alert SDRs when a lead’s score crosses a threshold. Option 2: Salesforce Step 1: Turn on Einstein Lead Scoring – Use AI score to score leads.
Use Flow Automation: Set up rules to assess leads according to their behaviors.
Add Product Analytics: If relevant for SaaS companies, integrate Amplitude or Mixpanel, so product usage can be tracked, thus updating the scores.
Set Dynamic Segments: Organize your leads based on their behavior and auto-enter them into nurturing tracks. Step 3: Move to Adaptive Lead Scoring Lead scores almost never stay static and decay over time. To keep your model effective: Decay Stale Leads: Subtract points from leads that haven’t acted in over 30 days (change based on longer sales processes).
Use Action-Weighted Points: High-intent actions (like demo requests or product searches) are worth much more than email opens.
Monitor and Optimize Monthly: Run CRM reports to ensure high-scoring leads are converting. Step 4: Align Sales & Marketing With Dynamic Scoring Your lead scoring model must work seamlessly with your sales team’s workflow. Here’s how: Define a “Marketing Qualified” Score (i.e., leads above 80 points trigger a follow-up).
Personalized Outreach: Sales should include recent high-intent actions in e-mails (e.g., “I noticed you read our enterprise security case); Let's chat!").
Feedback Loop: If sales report back that high-scoring leads are not converting, revise scoring weights accordingly. Conclusion If your lead scoring model is failing to deliver the results your sales team needs, it’s probably too rigid for today’s buyers. Today’s buyers don’t take a linear path — they research on their own, engage in bursts, and decide quickly. Traditional scoring models simply can’t keep pace with that kind of activity. Choose dynamic, intent-based lead scoring so your sales team only focuses on those most likely to convert. It moves past clunky lead scoring and instead into the realm of real-time engagement and behavioral signals. So, where do you start? Assess your existing model: What’s working, and what’s not? 🔝 Target intent signals: Look to buyers with clear readiness signals. ✅ Maintain updated scores: Keep your CRM up to mark with the latest information. ✅ Always iterate: Refine the model with feedback from your sales team. The shift will not, of course, happen overnight, but even small tweaks can dramatically impact both conversions and pipeline velocity. A well-implemented lead scoring model can thus work for you, not against you. It’s time to have a look at lead scoring and get it to work for your business in a smarter way. Would you like to take a stab at answering some of these questions? The link for the template is HERE. Interested in reading the content from all of our writing prompts? Click HERE. You have a lead scoring model in your CRM that just simply is not working? The harsh truth is that your lead scoring system as it stands, is probably broken. You are assigning points for job titles, for the number of times people visit your website, or open your emails, but when your sales team goes to these so-called “hot leads,” a lot of these people are not even interested. Sound familiar? The problem is that traditional lead scoring systems are based on the assumption that buyers go on a simple, linear journey. The reality is that today’s buyers are much more complex. By the time a buyer needs a solution, the decision is often already made. the decision is often already made. They do their homework long before you can even pronounce their name—glossing over reviews, measuring competitors, exploring industry blogs.


Their engagement is inconsistent — they may view a lot of content one week and none the next.


They may not respond to your emails, but engage with your product often — a clear sign of intent missed by static scoring models. They do their homework long before you can even pronounce their name—glossing over reviews, measuring competitors, exploring industry blogs. Their engagement is inconsistent — they may view a lot of content one week and none the next. They may not respond to your emails, but engage with your product often — a clear sign of intent missed by static scoring models. They do their homework long before you can even pronounce their name—glossing over reviews, measuring competitors, exploring industry blogs. They do their homework long before you can even pronounce their name—glossing over reviews, measuring competitors, exploring industry blogs. Their engagement is inconsistent — they may view a lot of content one week and none the next. Their engagement is inconsistent — they may view a lot of content one week and none the next. They may not respond to your emails, but engage with your product often — a clear sign of intent missed by static scoring models. They may not respond to your emails, but engage with your product often — a clear sign of intent missed by static scoring models. If you're still relying on static lead scoring, you might be missing out on valuable opportunities. This is where Dynamic Lead Scoring comes in - a more innovative, real-time approach that prioritizes leads based on what buyers do and intent signals. Dynamic Lead Scoring This guide will give you some ideas for how to create a dynamic lead-scoring model in your CRM. Then, by using AI, behavioral data, and automation, you can make sure your sales team is spending its time on the leads that are most likely to convert. Ready to get started? Let’s dive in. 🚀 Static and Dynamic Lead Scoring: What’s the Difference? Feature

Static Lead Scoring

Dynamic Lead Scoring



Data Sources

Demographics, firmographics, past actions

Real-time engagement, behavioral patterns, AI insights



Updates

Manual, periodic

Continuous, automated



Intent Signals

Clicks, form fills

Product usage, time-sensitive actions, in-market signals



Sales Alignment

Often misaligned

Predictive and relevant Feature

Static Lead Scoring

Dynamic Lead Scoring Data Sources

Demographics, firmographics, past actions

Real-time engagement, behavioral patterns, AI insights Updates

Manual, periodic

Continuous, automated Intent Signals

Clicks, form fills

Product usage, time-sensitive actions, in-market signals Sales Alignment

Often misaligned

Predictive and relevant Feature Static Lead Scoring Dynamic Lead Scoring Feature Feature Static Lead Scoring Static Lead Scoring Static Lead Scoring Dynamic Lead Scoring Dynamic Lead Scoring Dynamic Lead Scoring Data Sources Demographics, firmographics, past actions Real-time engagement, behavioral patterns, AI insights Data Sources Data Sources Data Sources Demographics, firmographics, past actions Demographics, firmographics, past actions Real-time engagement, behavioral patterns, AI insights Real-time engagement, behavioral patterns, AI insights Updates Manual, periodic Continuous, automated Updates Updates Updates Manual, periodic Manual, periodic Continuous, automated Continuous, automated Intent Signals Clicks, form fills Product usage, time-sensitive actions, in-market signals Intent Signals Intent Signals Intent Signals Clicks, form fills Clicks, form fills Product usage, time-sensitive actions, in-market signals Product usage, time-sensitive actions, in-market signals Sales Alignment Often misaligned Predictive and relevant Sales Alignment Sales Alignment Sales Alignment Often misaligned Often misaligned Predictive and relevant Predictive and relevant Static models make fixed point assignments (i.e., “10 points for an email open”), whereas dynamic scoring updates in real-time based on actual buyer behavior, providing more relevant and actionable insights. Step 1: Identify Your Dynamic Lead Scoring Model When creating a dynamic scoring model, there are three main types of data that you need to focus on: Fit Score (Who They Are)
IndustryCompany size
Job title
Company Revenue Fit Score (Who They Are) Industry Company size Job title Company Revenue Fit Score (Who They Are) Fit Score (Who They Are) Industry Company size Job title Company Revenue 👉 CRM Setup Tip: If your CRM is missing company data, identify gaps and use enrichment tools to fill them. Next, build workflows to move this data into contacts and create baseline scores and segments against your ICP (Ideal Customer Profile). Engagement Score (How They Interact With You)
Visits to your website (e.g. product or pricing pages = high intent)Content downloads
Webinar attendance
Email interactions
Social media engagement Engagement Score (How They Interact With You) Visits to your website (e.g. product or pricing pages = high intent) Content downloads Webinar attendance Email interactions Social media engagement Engagement Score (How They Interact With You) Engagement Score (How They Interact With You) Visits to your website (e.g. product or pricing pages = high intent) Content downloads Webinar attendance Email interactions Social media engagement 👉 CRM Setup Strategy: Use a marketing automation tool like HubSpot, Marketo, Pardot, etc., to track engagement, and automatically update scores dynamically. Intent Score (How Ready they are to Buy)
Engagement on the product (especially for SaaS companies)Frequent website visits
Competitor comparisons (activity on G2 or Capterra, etc.)
High engagement with sales materials (case studies, ROI calculators, etc.)
More web searches aligning with your product Intent Score (How Ready they are to Buy) Engagement on the product (especially for SaaS companies) Frequent website visits Competitor comparisons (activity on G2 or Capterra, etc.) High engagement with sales materials (case studies, ROI calculators, etc.) More web searches aligning with your product Intent Score (How Ready they are to Buy) Intent Score (How Ready they are to Buy) Engagement on the product (especially for SaaS companies) Frequent website visits Competitor comparisons (activity on G2 or Capterra, etc.) High engagement with sales materials (case studies, ROI calculators, etc.) More web searches aligning with your product 👉 CRM Setup Hack: Link intent data solutions, such as HubSpot Breeze, 6sense, Bombora, or others to your CRM. Step 2: Add Scoring Automation in Your CRM Option 1: HubSpot 3rd Party Data Enrichment: Use ZoomInfo, HubSpot Breeze, Bombora, etc. to track buyer intent outside of your own channels. Establish Scoring Models – Fit, Engagement, and Intent scores should each have their own models. Sales Alerts: Automatically alert SDRs when a lead’s score crosses a threshold. 3rd Party Data Enrichment: Use ZoomInfo, HubSpot Breeze, Bombora, etc. to track buyer intent outside of your own channels. 3rd Party Data Enrichment: Use ZoomInfo, HubSpot Breeze, Bombora, etc. to track buyer intent outside of your own channels. Establish Scoring Models – Fit, Engagement, and Intent scores should each have their own models. Establish Scoring Models – Fit, Engagement, and Intent scores should each have their own models. Sales Alerts: Automatically alert SDRs when a lead’s score crosses a threshold. Sales Alerts: Automatically alert SDRs when a lead’s score crosses a threshold. Option 2: Salesforce Step 1: Turn on Einstein Lead Scoring – Use AI score to score leads. Use Flow Automation: Set up rules to assess leads according to their behaviors. Add Product Analytics: If relevant for SaaS companies, integrate Amplitude or Mixpanel, so product usage can be tracked, thus updating the scores. Set Dynamic Segments: Organize your leads based on their behavior and auto-enter them into nurturing tracks. Step 1: Turn on Einstein Lead Scoring – Use AI score to score leads. Use Flow Automation: Set up rules to assess leads according to their behaviors. Add Product Analytics: If relevant for SaaS companies, integrate Amplitude or Mixpanel, so product usage can be tracked, thus updating the scores. Set Dynamic Segments: Organize your leads based on their behavior and auto-enter them into nurturing tracks. Step 3: Move to Adaptive Lead Scoring Lead scores almost never stay static and decay over time. To keep your model effective: Decay Stale Leads: Subtract points from leads that haven’t acted in over 30 days (change based on longer sales processes). Use Action-Weighted Points: High-intent actions (like demo requests or product searches) are worth much more than email opens. Monitor and Optimize Monthly: Run CRM reports to ensure high-scoring leads are converting. Decay Stale Leads: Subtract points from leads that haven’t acted in over 30 days (change based on longer sales processes). Use Action-Weighted Points: High-intent actions (like demo requests or product searches) are worth much more than email opens. Monitor and Optimize Monthly: Run CRM reports to ensure high-scoring leads are converting. Step 4: Align Sales & Marketing With Dynamic Scoring Your lead scoring model must work seamlessly with your sales team’s workflow. Here’s how: Define a “Marketing Qualified” Score (i.e., leads above 80 points trigger a follow-up). Personalized Outreach: Sales should include recent high-intent actions in e-mails (e.g., “I noticed you read our enterprise security case); Let's chat!"). Feedback Loop: If sales report back that high-scoring leads are not converting, revise scoring weights accordingly. Define a “Marketing Qualified” Score (i.e., leads above 80 points trigger a follow-up). Personalized Outreach: Sales should include recent high-intent actions in e-mails (e.g., “I noticed you read our enterprise security case); Let's chat!"). Feedback Loop: If sales report back that high-scoring leads are not converting, revise scoring weights accordingly. Conclusion If your lead scoring model is failing to deliver the results your sales team needs, it’s probably too rigid for today’s buyers. Today’s buyers don’t take a linear path — they research on their own, engage in bursts, and decide quickly. Traditional scoring models simply can’t keep pace with that kind of activity. Choose dynamic, intent-based lead scoring so your sales team only focuses on those most likely to convert. It moves past clunky lead scoring and instead into the realm of real-time engagement and behavioral signals. So, where do you start? Assess your existing model: What’s working, and what’s not? 🔝 Target intent signals: Look to buyers with clear readiness signals. ✅ Maintain updated scores: Keep your CRM up to mark with the latest information. ✅ Always iterate: Refine the model with feedback from your sales team. The shift will not, of course, happen overnight, but even small tweaks can dramatically impact both conversions and pipeline velocity. A well-implemented lead scoring model can thus work for you, not against you. It’s time to have a look at lead scoring and get it to work for your business in a smarter way. Would you like to take a stab at answering some of these questions? The link for the template is HERE. Interested in reading the content from all of our writing prompts? Click HERE. Would you like to take a stab at answering some of these questions? The link for the template is HERE . Interested in reading the content from all of our writing prompts? Click HERE . HERE HERE HERE HERE

Stories, News & Mentions about AMPLITUDE

Categories

Trending Topics

Stories, News & Mentions about AMPLITUDE

Amplitude’s Stories on HackerNoon

Amplitude's News & Mentions

Categories

Trending Topics